Service Attachment — Artificial Intelligence Services

This Service Attachment is between We Solve Problems, LLC (“Provider,” “we,” “us”) and the client identified on the applicable Order (“Client,” “you”). Together with the Order, the Services Agreement, and other applicable Service Attachments it forms the Agreement between the parties.

We deliver only the Services itemized on the Order. Additional services require a new Order.

1. Definitions

“AI Services” means the services itemized on the Order under this Attachment, which may include AI strategy consulting, solution design, implementation, automation development, AI application development, and ongoing operation of an AI system.

“AI Tool” means a third-party product with artificial intelligence capability, whether sold as an AI product or as an AI feature within a broader product.

“AI-Generated Output” means any content, data, analysis, code, or other material produced by an AI Tool or AI System in the course of the AI Services.

“AI System” means a configured assembly of one or more AI Tools, prompts, connections, credentials, and automations that we design, build, or operate for Client under an Order.

“Agentic Capability” means the ability of an AI System to take actions in Client’s environment — writing files, sending messages, executing commands, calling APIs, or modifying data — rather than only producing text or analysis for a person to read.

“Client AI Data” means data Client provides to us, or makes reachable by an AI System, for the purpose of the AI Services.

2. What is included and what is billable

In plain terms. Recommending a tool is free. Building with it is paid. This is the same line we draw with accounting software: we will tell you which package fits and help you get accounts set up, but we do not become your bookkeeper.

2.1 Included in managed services at no additional charge

Where Client subscribes to managed services, the following are included and are not AI Services under this Attachment:

naming AI Tools that may suit Client’s stated needs, and our general opinion of them;
provisioning user accounts, licenses, and single sign-on for an AI Tool, on the same basis as any other application;
applying tenant-level administrative settings an AI Tool exposes;
answering general questions about what an AI Tool does; and
incidental guidance during a support interaction.

2.2 AI Services, billable under an Order

The following are AI Services and require an Order:

assessing Client’s readiness, data, or processes for AI adoption, and documenting findings;
designing an AI System, including tool selection for a specific workflow, architecture, and data flow;
building, configuring, or deploying an AI System, including prompts, automations, integrations, and connections to Client systems;
developing custom AI applications, agents, chatbots, or voice assistants;
operating, monitoring, or maintaining an AI System after deployment;
structured training or workshops delivered to Client’s personnel; and
any work that grants an AI Tool access to Client data or systems beyond a single user’s own access.

2.3 Why the line is drawn here

Client acknowledges that work performed without charge may fall outside the definition of professional services in our errors-and-omissions policy, and that the allocation above reflects that. Where Client wants work performed that section 2.1 does not include, it is available under an Order.

2.4 Training

Structured training is an AI Service under section 2.2 and requires an Order that describes what is delivered. We do not deliver structured training on an unbilled basis.

2a. Cadence and response

Unless the Order states otherwise, an AI Services retainer includes:

two (2) working sessions per calendar month with Client’s leadership, scheduled by agreement;
one (1) hands-on enablement session per calendar month, up to ninety (90) minutes, on a topic from the agreed menu; and
asynchronous decision support by email and chat at reasonable volume.

We will respond to an asynchronous request within two (2) Business Days. This is an advisory engagement, not a help desk and not an on-call service; where Client also buys Managed IT Services, the four-Business-Hour target in section 19.1 governs those.

Unused sessions do not carry forward. Additional sessions are billed separately or folded into a build engagement.

3. Independent judgment and no reliance

In plain terms. When we recommend an AI tool, that is our opinion, not a promise that it is safe or legal for what you want to do with it. You decide, and you check the output before you act on it.

3.1 Recommendations are opinions

Our recommendations regarding AI Tools, vendors, models, configurations, and use cases are professional opinions offered as one or more reasonable options, and are not representations of fact regarding suitability, safety, accuracy, security, or regulatory compliance.

Client exercises its own independent business judgment in selecting, approving, and using any AI Tool, and is solely responsible for reviewing and validating all AI-Generated Output before relying on it or acting on it.

3.2 No regulatory determination

Unless an Order expressly states otherwise, we have not evaluated and do not represent that any AI Tool is suitable for Client’s regulated data or for use in any regulated activity.

We do not provide legal, accounting, medical, or regulatory advice. No statement by us constitutes a legal opinion, certification, attestation, or determination that Client is compliant with any law, regulation, or standard, including any law governing artificial intelligence.

3.3 Ownership of policy and compliance

We set up and guide AI governance, data-handling practice, and vendor-risk posture. Ownership of policy, risk, and regulatory compliance remains with Client’s leadership, on every engagement.

3.4 Human review

Client is responsible for reviewing AI-Generated Output before it is used in a decision, a filing, a communication to a third party, or a transaction. Client will not represent AI-Generated Output as reviewed by us unless an Order says we review it.

4. High-impact uses

In plain terms. If you want to use AI to make decisions about people — hiring, lending, insurance, housing, health, education — that is a different category with real regulation attached, and it needs to be written into an Order before it starts.

4.1 Human in the loop

Client will not use an AI System we build or operate to make or substantially inform a Consequential Decision without meaningful human review of each output before it takes effect.

A “Consequential Decision” is one producing a legal or similarly significant effect on an individual’s access to or terms of employment, education, lending or credit, housing, insurance, healthcare, or an essential government service.

4.2 Disclosure required

Client will notify us in writing before using an AI System for a Consequential Decision. Any such use requires an Order that identifies it, and may require changes to the AI System, additional documentation, or a Data Processing Agreement.

4.3 If Client proceeds without disclosure

Where Client uses an AI System for a Consequential Decision without the Order required by section 4.2, Client is responsible for the consequences, and we may suspend the affected Services under section 8.

5. Approved tools and change control

In plain terms. We agree in writing which AI tools are in play. If a vendor switches on an AI feature inside a product you already use, that is a change, and we deal with it rather than discovering it later.

5.1 Approved tool list

The AI Tools we will deploy, configure, or operate for Client are those identified on the Order or on an approved-tool list the parties maintain in writing. We will not introduce a new AI Tool into Client’s environment without Client’s written approval, which may be given by email.

5.2 Vendor-introduced AI features

Where a vendor enables AI capability inside a product already deployed in Client’s environment, we will notify Client within a commercially reasonable time after we become aware of it and recommend whether to enable, restrict, or disable it. Client decides. Absent Client’s instruction, we will apply the more restrictive setting the product supports.

5.3 Model and version changes

AI Tool vendors change underlying models, capabilities, and behavior without notice. We are not responsible for a change in AI System behavior caused by a vendor’s change to its model, and remediation is billable under a new Order unless the Order provides otherwise.

6. Third-party AI terms

6.1 Flow-down

AI Tools are Third-Party Services under the Services Agreement and the Schedule of Third-Party Services. Client’s rights in an AI Tool are governed by that vendor’s terms, which Client is responsible for reviewing and complying with, including its acceptable use policy.

6.2 Plan type and data handling

Client acknowledges that an AI Tool’s data-retention and training behavior depends on the plan under which it is licensed, and that consumer and individual plans frequently differ from business and enterprise plans in whether vendor personnel may review inputs and whether inputs are used for model training.

Where Client’s use involves Regulated Data, Client is responsible for licensing the AI Tool on a plan that supports the required contractual protections, including any vendor data processing agreement or business associate agreement. We will identify what the plan requires; procuring it is Client’s decision and cost.

6.3 No vendor warranty

We do not warrant any AI Tool, its availability, its output, or its continued existence. A vendor’s discontinuation of a product, change in terms, or restriction of access is not our breach.

7. Acceptable use

Client will not use, and will not permit its personnel to use, an AI System we build or operate:

in violation of any AI Tool vendor’s acceptable use policy or terms;
to generate content that is unlawful, defamatory, harassing, or infringing;
to process Regulated Data outside the terms of an Order and Data Processing Agreement;
to make a Consequential Decision except as section 4 permits;
to impersonate a person without disclosure where disclosure is required by law;
to circumvent a safety control, content filter, or usage limit in an AI Tool; or
for High-Risk Use, meaning any application where failure could lead to death, serious bodily injury, or severe physical or environmental damage.

Client will indemnify us against third-party claims arising from its breach of this section.

8. Suspension

We may suspend the AI Services immediately on written notice where we reasonably believe Client is using an AI System in breach of section 4 or section 7, or in a way that creates a risk of imminent harm, legal liability, or breach of a vendor’s terms.

We will describe the basis for the suspension and restore the Services promptly once the cause is resolved. Suspension under this section does not relieve Client of its payment obligations for Services already delivered.

9. Agentic capability and credentials

In plain terms. An AI tool that can take actions in your systems is a different risk from one that writes text. Where we build one, there are rules about what it can reach.

9.1 Scope of access

Where an AI System has Agentic Capability, the Order will identify the systems it may reach and the actions it may take. We will configure it to the least privilege that allows it to do the job described.

9.2 Production credentials

We will not configure an AI System with a credential that grants unattended write or delete access to a production system, a backup, or a payment system unless the Order expressly authorizes it and identifies the compensating controls.

9.3 Client-initiated changes

Where Client or a third party changes an AI System’s permissions, credentials, prompts, or connections after we deliver it, we are not responsible for the consequences, and remediation is billable.

10. Our obligations

10.1 Standard of performance

We will perform the AI Services in a professional and workmanlike manner, consistent with the standard of care ordinarily exercised by managed service providers performing comparable work. This is the only warranty we give in respect of the AI Services.

10.2 Security

We will apply commercially reasonable administrative, technical, and physical safeguards to Client AI Data in our possession, consistent with the Services Agreement and any Data Processing Agreement.

10.3 Notification

We will notify Client promptly after becoming aware of a security incident affecting Client AI Data in our possession, in accordance with the Services Agreement and any Data Processing Agreement.

10.4 Documentation

For each AI System we deploy, we will provide Client with a written record identifying the AI Tools used, the data the system can reach, the actions it can take, the accounts and credentials involved, and who to contact at the vendor. We will update it when we change the system.

11. Client obligations

Client will:

provide accurate and complete information about its systems, data, and objectives, and tell us promptly when they change;
provide the access and cooperation the AI Services require;
ensure it has the right to provide us with the data it provides, and that providing it complies with applicable privacy law and any obligation Client owes a third party;
ensure the data meets the quality standards the Order identifies;
comply with the laws applicable to Client and Client’s industry, including any law governing artificial intelligence;
maintain the confidentiality of our materials and methods; and
notify us promptly of any malfunction, unexpected behavior, or incident involving an AI System.

12. Acceptance testing

12.1 Notification and period

On completion of a deployment or custom development, we will notify Client that it is ready for acceptance testing. Client has fifteen (15) business days from that notice to test. We will provide reasonable support during the period, including documentation and technical assistance.

12.2 Criteria

The AI Services are accepted if they perform in accordance with the specifications on the Order and no critical defect is identified that would materially impair Client’s ability to use them for their intended purpose. Minor defects are documented and scheduled for correction and do not prevent acceptance.

Client acknowledges that AI-Generated Output is probabilistic and that variation in output between runs is not, by itself, a defect.

12.3 Failure

If Client determines the AI Services have not passed, Client will notify us in writing within ten (10) business days of the end of the testing period, describing the deficiencies. We then have thirty (30) business days to correct them and resubmit.

If the AI Services fail to meet the acceptance criteria after three attempts, Client may either reject them and terminate this Attachment as to the failed services without penalty, receiving a refund of fees paid for those services, or accept them with the deficiencies subject to a negotiated fee reduction.

12.4 Deemed acceptance

If Client does not deliver a written notice under section 12.3 within the period stated, the AI Services are deemed accepted.

13. Intellectual property

In plain terms. You own your data and the outputs. We own our methods and the tooling we bring. What we build specifically for you is yours once you have paid for it.

13.1 Pre-existing property

Each party retains its pre-existing intellectual property. Client grants us a non-exclusive, royalty-free license to use Client’s pre-existing intellectual property solely to perform the AI Services.

13.2 AI-Generated Output

As between the parties, Client owns the AI-Generated Output produced for Client, subject to any rights of the AI Tool vendor in the underlying model and to the vendor’s terms.

Client is responsible for ensuring its use of AI-Generated Output complies with applicable law and does not infringe a third party’s rights. Client acknowledges that the copyright status of AI-generated material is unsettled and that we make no representation that any AI-Generated Output is protectable or free of third-party claims.

13.3 Custom development

Custom AI systems, applications, agents, and prompts we create specifically for Client under an Order are Client’s on payment in full for them, excluding our pre-existing materials incorporated into them, for which we grant Client the license in section 5.3 of the Services Agreement.

We retain the right to use the general knowledge, skills, techniques, and experience gained in performing the AI Services.

13.4 Third-party and open-source components

We may use third-party and open-source components in delivering the AI Services. We will comply with their licenses and will tell Client where a component imposes an attribution requirement or a restriction on the use of output.

13.5 Aggregated data

We may use aggregated and de-identified operational data derived from the AI Services for analytics, benchmarking, and improving our own delivery, provided it does not identify Client, its personnel, or its customers, and does not include Client’s content. We will not submit Client AI Data to an AI Tool vendor for the purpose of training that vendor’s models.

14. Exclusions

We are not responsible for failures or outcomes caused by:

a third-party service or product not supplied or controlled by us, including an AI Tool vendor’s outage, deprecation, or change in terms;
Client’s disregard of our recommendations or instructions;
modification of an AI System by Client or a third party;
inaccuracies, fabrications, or omissions in AI-Generated Output that human review would reasonably have caught;
bias or unfairness inherent in a third-party model or in data Client supplied;
the absence of a detailed explanation for a model’s decision;
inadequate or poor-quality data supplied by Client;
a change in an AI System’s behavior caused by a vendor’s model update or continuous learning;
a third-party claim of infringement arising from AI-Generated Output;
a change in law or regulation restricting the use of AI;
prompt injection or a comparable attack on an AI Tool that its vendor’s controls failed to prevent; or
a force majeure event as described in the Services Agreement.

None of the above limits our obligations under section 10.1.

15. Disclaimer

In plain terms. AI is not deterministic and cannot be warranted the way a server can. Beyond performing our work professionally, we make no promises about what the AI produces.

EXCEPT FOR THE STANDARD OF PERFORMANCE IN SECTION 10.1, THE AI SERVICES, AI TOOLS, AI SYSTEMS, AND AI-GENERATED OUTPUT ARE PROVIDED “AS IS” AND “AS AVAILABLE.” WE DISCLAIM ALL OTHER WARRANTIES, EXPRESS, IMPLIED, OR STATUTORY, INCLUDING MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, AND NON-INFRINGEMENT.

WE MAKE NO WARRANTY THAT THE AI SERVICES WILL MEET CLIENT’S REQUIREMENTS OR ACHIEVE ANY INTENDED RESULT, OR THAT AI-GENERATED OUTPUT WILL BE ACCURATE, COMPLETE, OR RELIABLE. CLIENT IS RESPONSIBLE FOR SELECTING THE AI SERVICES IT BELIEVES WILL ACHIEVE ITS INTENDED RESULTS AND FOR REVIEWING AND VALIDATING THE OUTPUT.

Liability under this Attachment is governed by section 11 of the Services Agreement.

16. Term and termination

16.1 Term

This Attachment begins on the Service Start Date on the Order and runs for the initial term stated there.

16.2 Renewal

THIS ATTACHMENT RENEWS AUTOMATICALLY FOR SUCCESSIVE TWELVE (12) MONTH TERMS UNLESS EITHER PARTY GIVES WRITTEN NOTICE OF NON-RENEWAL AT LEAST THIRTY (30) DAYS BEFORE THE END OF THE THEN-CURRENT TERM. We will send Client a reminder of the renewal date and the notice deadline at least sixty (60) days before the end of each term. Renewal is at our then-current rates, subject to section 3.2 of the Services Agreement.

16.3 Month-to-month

If the Order states no initial term, we deliver the Services month to month until either party gives written notice, ending at the close of the following calendar month.

16.4 Termination by Client for cause

Client may terminate this Attachment on thirty (30) days’ written notice if we materially fail to meet our obligations and do not cure within thirty (30) days of written notice, or if we cease business operations.

16.5 Termination by Client without cause

After ninety (90) days from the Service Start Date, Client may terminate this Attachment without cause on sixty (60) days’ written notice, paying a termination fee equal to fifty percent (50%) of the recurring monthly fees remaining through the end of the then-current term at the prices then in effect.

16.6 Termination by us

We may terminate this Attachment on sixty (60) days’ written notice, or immediately for Client’s unlawful or abusive conduct, or for a breach of section 7 that Client does not cure within ten (10) days of notice.

16.7 Effect of termination

On termination, and provided Client is current on amounts owed, we will assist in the orderly transition of the AI Services at our then-prevailing rates under a separate Order, and will deliver the documentation described in section 10.4, a copy of Client’s prompts and configurations, and a copy of Client AI Data in a commercially reasonable format.

Client is responsible for third-party AI Tool fees and any vendor termination charges.